Method of extracting contour of image, method of extracting object from image, and video transmission system using the same method

Abstract

An object extraction method comprising the steps of comparing each portion of first image data obtained by capturing an extraction object with another portion of the first image data or second image data, thereby generating shape data as an image representing a provisional region of the extraction object in the first image data, and matching a contour of the shape data to a contour of the extraction object by using the first image data and the provisionally generated shape data.

1 - 8 . (canceled)  
   
   
       9 . An image contour extraction method comprising: 
 displaying an original image on a display screen based on original image data;    manually setting a rough contour of an object displayed on the display screen and generating shape data corresponding to the rough contour; and    performing contour position correction by a self-affine mapping technique to match an actual contour of the object to the rough contour, the contour position correction including:    performing reduction processing on the original image data and shape data to obtain a reduced original image and a reduced rough contour,    performing coarse contour position correction by the self-affine mapping technique using the reduced original image data and the reduced rough contour, coarsely to bring the rough contour close to the contour of the original image,    performing main contour position correction by the self-affine mapping technique using the coarsely adjusted shape data and original image data obtained by the coarse contour position correction, and    matching the rough contour to the contour of the original image.    
   
   
       10 . A method according to  claim 9 , wherein the coarse contour position correction comprises arranging a plurality of search reference blocks, each having a first size, on the reduced rough contour, searching, for each of the search reference blocks, the reduced image data corresponding to the search reference block for a similar block having a texture similar to that of the reduced original image and a second size larger than the first size of the search reference block, and replacing all the search reference blocks with correction shape data obtained by extracting data of the similar blocks from the shape data and reducing the data to the first size.  
   
   
       11 . A method according to  claim 10 , wherein the main contour position correction comprises arranging a plurality of search reference blocks, each having a third size, on the rough contour, searching, for each of the search reference blocks, the original image data corresponding to the search reference block for a similar block having a texture similar to that of the original image and a fourth size larger than the third size of the search reference block, and replacing all the search reference blocks with correction shape data obtained by extracting data of the similar blocks from the rough contour and reducing the data to the third size.  
   
   
       12 . A method according to  claim 9 , wherein the coarse contour position correction comprises obtaining a ½ reduced original image by reducing the original image and rough contour to ½.  
   
   
       13 . A method according to  claim 9 , wherein the coarse contour position correction is repeated a plurality of number of times while a size of the search reference block is decreased every coarse contour position correction.  
   
   
       14 . A method according to  claim 9 , wherein the coarse contour position correction comprises limiting a similar block search range to a direction perpendicular to a direction of a contour of shape data in a search reference block.  
   
   
       15 . A method according to  claim 9 , wherein the coarse contour position correction comprises reducing a block size of a search reference block to ½ vertically and horizontally, checking whether the reduced block size of the search reference block is smaller than a reference value, and advancing the main contour position correction if the block size is smaller than the reference value and returning to the reduction step if the block size is not smaller than the reference value.  
   
   
       16 - 17 . (canceled)  
   
   
       18 . A computer system comprising: 
 means for displaying an original image on a display screen based on original image data;    means for manually setting a rough contour of an object displayed on the display screen and generating shape data corresponding to the rough contour; and    means for performing contour position correction by a self-affine mapping technique to match an actual contour of the object to the rough contour, the contour position correction including:    performing reduction processing on the original image data and shape data to obtain a reduced original image and a reduced rough contour, and    performing coarse contour position correction by the self-affine mapping technique using the reduced original image data and the reduced rough contour, coarsely to bring the rough contour close to the contour of the original image,    means for performing main contour position correction by the self-affine mapping technique using the coarsely adjusted shape data and original image data obtained by the coarse contour position correction, and    means for matching the rough contour to the contour of the original image.    
   
   
       19 . The computer system according to  claim 18 , wherein the coarse contour position correction comprises arranging a plurality of search reference blocks, each having a first size, on the reduced rough contour, searching, for each of the search reference blocks, the reduced image data corresponding to the search reference block for a similar block having a texture similar to that of the reduced original image and a second size larger than the first size of the search reference block, and replacing all the search reference blocks with correction shape data obtained by extracting data of the similar blocks from the shape data and reducing the data to the first size.  
   
   
       20 . The computer system according to  claim 18 , wherein the main contour position correction comprises arranging a plurality of search reference blocks, each having a third size, on the rough contour, searching, for each of the search reference blocks, the original image data corresponding to the search reference block for a similar block having a texture similar to that of the original image and a fourth size larger than the third size of the search reference block, and replacing all the search reference blocks with correction shape data obtained by extracting data of the similar blocks from the rough contour and reducing the data to the third size.  
   
   
       21 . The computer system according to  claim 18 , wherein the coarse contour position correction comprises obtaining a ½ reduced original image by reducing the original image and rough contour to ½.  
   
   
       22 . The computer system according to  claim 18 , further including repeating the coarse contour position correction a plurality of number of times while a size of the search reference block is decreased every coarse contour position correction.  
   
   
       23 . The computer system according to  claim 18 , wherein the coarse contour position correction comprises limiting a similar block search range to a direction perpendicular to a direction of a contour of shape data in a search reference block.  
   
   
       24 . The computer system according to  claim 18 , wherein the coarse contour position correction comprises reducing a block size of a search reference block to ½ vertically and horizontally, checking whether the reduced block size of the search reference block is smaller than a reference value, and advancing the main contour position correction if the block size is smaller than the reference value and returning to the reduction step if the block size is not smaller than the reference value.